For the most part even the best behavioral interventions for prevention and treatment of drug abuse and related conditions such as HIV/AIDS and Hepatitis C have not been systematically optimized to achieve their maximum potency and public health benefit. Collins et al. (in press) have outlined a framework for systematic optimization of behavioral interventions. This framework, called the Multiphase Optimization STrategy (MOST), divides behavioral intervention research into three phases: screening, in which a set of intervention components under consideration is efficiently sorted through to select the promising candidates for further investigation;refining, in which optimal intervention doses are determined and interactions between program components and characteristics of the individual, environment, and so on are examined;and confirming, in which an optimized program constructed on the basis of the results of the screening and refining phases is subjected to a full intervention trial. MOST relies on specialized, very economical designs that are commonly used in engineering, to allow the user to address research questions of primary importance using the smallest possible number of experimental conditions. Using computer simulations, the present project will carry out an investigation of MOST, based on ongoing input from drug abuse intervention scientists. The simulations will investigate the effects of key aspects of drug abuse intervention research that occur commonly in empirical settings, to establish how a high degree of accuracy in decision making can be maintained in MOST under noisy field research conditions. They will also explore how economic cost-effectiveness analysis can be incorporated into the MOST procedure. The end result will be dissemination of clear guidelines for intervention scientists who wish to use MOST in their work. This will lead to more potent behavioral interventions, which will have a larger and more sustained impact on the reduction of drug abuse and related morbidity and mortality.